The binding between stimuli and responses has been proved an important mechanism in action control. It has been found that features of a stimulus and an executed response are integrated together into one event file. A re-encounter with one or more of the stored features causes the automatic retrieval of the previously constructed event file including the previous integrated response (Hommel, 2004). The distractor-response binding effect evidenced that even irrelevant stimuli can be integrated with a response and thereby have an impact on behavior. In the present study, we focused on the level of representation of response-irrelevant distractor stimuli. To this end, a crossmodal distractor-response binding paradigm was used that enables the differentiation between the perceptual and conceptual representation of the distractor by allowing the systematic variation of conceptual distractor features independent of perceptual repetitions. The results suggest that the repetition of perceptual distractor features is indispensable for the initiation of the retrieval process while the sole repetition of conceptual distractor features is not sufficient to start the retrieval process.